FakeDetect / README.md
Basti1110's picture
Update README.md
db46f4e verified
metadata
title: FakeDetect
emoji: 🐠
colorFrom: purple
colorTo: red
sdk: docker
pinned: false
license: mit
short_description: A minimal Node.js + Gemini-powered web app to analyze text.

πŸ•΅οΈ FakeDetect - OSINT Intelligence Analyzer

FakeDetect is a real-time Open Source Intelligence (OSINT) web application that uses Google Gemini AI to analyze and visualize multilingual social media and news content.

✨ Features

  • 🧠 AI-powered content interpretation using Gemini 1.5 Flash
  • πŸ“Š Sentiment analysis (positive / neutral / negative breakdown)
  • 🌐 Network influence graph for user/entity clustering
  • πŸ“Œ Structured report output (JSON + HTML)
  • πŸ’Ύ Exportable report as modern dark-mode HTML
  • πŸ’¬ Multilingual support (EN, DE, ES, FR, RU)

πŸ”§ Tech Stack

  • Frontend: HTML + TailwindCSS
  • Backend: Node.js (Express) + Python (matplotlib, networkx)
  • AI: Google Generative AI (Gemini)
  • Deployment: Hugging Face Spaces (Docker SDK)

πŸš€ Quick Start (Docker)

git clone https://github.com/Basti-devel/fakedetect.git
cd fakedetect

# Build the container
docker build -t fakedetect .

# Run the app
docker run -p 7860:7860 fakedetect

Access the app at: http://localhost:7860

πŸ“ Project Structure

β”œβ”€β”€ app.js                  # Node.js backend (Express)
β”œβ”€β”€ render_report.py        # Python rendering (charts + HTML)
β”œβ”€β”€ index.html              # Frontend interface
β”œβ”€β”€ osint_dark_template.html # Dark HTML template for reports
β”œβ”€β”€ requirements.txt        # Python dependencies
β”œβ”€β”€ package.json            # Node dependencies
β”œβ”€β”€ space.yaml              # Hugging Face Space config
β”œβ”€β”€ Dockerfile              # Unified deployment container
└── /tmp                    # Runtime output directory

πŸ›‘ Ethics & Compliance

FakeDetect adheres to:

  • πŸ“ GDPR and digital privacy principles
  • πŸ•ŠοΈ Journalistic ethics and OSINT standards
  • πŸ›‘ Avoids speculative conclusions, hate speech, or bias

🧠 Sample Use Cases

  • πŸ—ž Investigative journalism
  • πŸ›‘ NGO threat reporting
  • 🧡 Telegram/Reddit sentiment tracking
  • 🧾 Misinformation early detection

© 2025 FakeDetect by Sebastian Friedrich Nestler. Built with ❀️ for responsible intelligence analysis.